Today, predicting the behavior of consumers with available statistical methods is ineffective. This is in part because, siloed, device-specific approaches are currently being utilized by advertisers and advertising agencies. These methods limit the information available at the individual consumer level for lack of an ability to track how consumers behave across all of their media devices. For example, advertisers receive data from TV panel companies such as Nielsen, and use the information to decide how they are going to design and implement an advertising campaign. Panel companies utilize a small group of (usually 15,000-20,000) people selected statistically to be representative of the population, and use statistical extrapolation from viewing data on the panel to make deductions about the population at large. Alternatively, the advertisers will receive data from online panels such as Comscore, Nielsen, and Kantar, which track where the audience is online. Cable TV operators sell their own viewership data from their subscribers. The advent of online technology has meant that it is possible to directly collect data on large numbers of consumers, with the potential to achieve to more accurate assessment of population viewing habits. Nevertheless, it has not been possible so far to design a single advertising campaign that addresses all of the different types of media at once because of the silo'ed nature of the data, and interface points to the various media conduits.
Thus, a direct, automated, aggregate view of consumer behavior does not currently exist today. Instead, advertisers and brand managers look at each data source separately. Human analysts guide the selection of advertising inventory based on, for example, Excel data tables and other static data management tools. This results in low selection efficiency and delays in responding to market trends. Consumers are not disparate silos of preference, yet the market for advertising treats them as such due to limitations in the available methods, most of which are incapable of quickly and accurately integrating information about how consumers behave across all of their devices.
Strategy for TV is planned according to TV-specific criteria, and web and mobile advertising, which include sub categories such as social media, are each planned separately. It is difficult to accurately predict consumer behavior, when data about their behavior is fragmented due to their multi-screen usage. Furthermore, consumers now have a large range of possible behaviors. One consumer might favor using a set top box to watch sports, but may prefer using a mobile device to watch YouTube. Another consumer may favor Hulu access via their desktop computer, but only accesses social media via their smart phone. There is no practical method for normalizing a complete view of consumer data to predict how and when they will access certain media devices throughout the day.
These complications have ramifications for advertisers in their efforts to both design effective campaigns and assess—quantitatively—the effectiveness of an ongoing campaign effort. The problem is especially acute for branding purposes, when it's more difficult for an advertiser to be confident that the right target audience saw a particular advertisement, or that the advertisement reached as many people in the target audience as had been intended.
There are other aspects that advertisers would like to know that cannot be reliably calculated today. For example, it is not possible to tell from Nielsen data whether the same consumer saw an advertisement on both TV and mobile. A panel company has to extrapolate and can only imperfectly estimate an answer to this question, if at all. Furthermore, panel companies simply report data and trends in data, but typically hold back from making a recommendation to an advertiser to tailor or target its advertising content differently.
The discussion of the background herein is included to explain the context of the technology. This is not to be taken as an admission that any of the material referred to was published, known, or part of the common general knowledge as at the priority date of any of the claims found appended hereto.
Throughout the description and claims of the application the word “comprise” and variations thereof, such as “comprising” and “comprises”, is not intended to exclude other additives, components, integers or steps.